CN116597917B - Molecular docking method and device based on light quantum computer - Google Patents
Molecular docking method and device based on light quantum computer Download PDFInfo
- Publication number
- CN116597917B CN116597917B CN202310882617.XA CN202310882617A CN116597917B CN 116597917 B CN116597917 B CN 116597917B CN 202310882617 A CN202310882617 A CN 202310882617A CN 116597917 B CN116597917 B CN 116597917B
- Authority
- CN
- China
- Prior art keywords
- ligand
- receptor
- grid
- docking
- match
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000003032 molecular docking Methods 0.000 title claims abstract description 67
- 238000000034 method Methods 0.000 title claims abstract description 62
- 239000003446 ligand Substances 0.000 claims abstract description 152
- 238000005457 optimization Methods 0.000 claims abstract description 38
- 238000010586 diagram Methods 0.000 claims abstract description 28
- 210000001503 joint Anatomy 0.000 claims abstract description 21
- 230000001427 coherent effect Effects 0.000 claims abstract description 11
- 102000005962 receptors Human genes 0.000 claims description 119
- 108020003175 receptors Proteins 0.000 claims description 119
- 230000003287 optical effect Effects 0.000 claims description 19
- 230000008569 process Effects 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 4
- 238000010168 coupling process Methods 0.000 claims description 4
- 238000005859 coupling reaction Methods 0.000 claims description 4
- 239000000243 solution Substances 0.000 abstract description 21
- 239000003814 drug Substances 0.000 abstract description 14
- 102000004169 proteins and genes Human genes 0.000 abstract description 11
- 108090000623 proteins and genes Proteins 0.000 abstract description 11
- 238000004422 calculation algorithm Methods 0.000 abstract description 5
- 229940079593 drug Drugs 0.000 abstract description 4
- 238000009510 drug design Methods 0.000 abstract description 4
- 238000000302 molecular modelling Methods 0.000 abstract description 2
- 239000012086 standard solution Substances 0.000 abstract description 2
- 230000001133 acceleration Effects 0.000 abstract 1
- 125000004429 atom Chemical group 0.000 description 45
- 230000006872 improvement Effects 0.000 description 7
- 150000003384 small molecules Chemical class 0.000 description 7
- 238000012827 research and development Methods 0.000 description 5
- 238000007877 drug screening Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 150000002611 lead compounds Chemical class 0.000 description 4
- 230000005366 Ising model Effects 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000012900 molecular simulation Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000009509 drug development Methods 0.000 description 2
- 230000005283 ground state Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000013178 mathematical model Methods 0.000 description 2
- 239000002096 quantum dot Substances 0.000 description 2
- 238000010845 search algorithm Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 102000012440 Acetylcholinesterase Human genes 0.000 description 1
- 108010022752 Acetylcholinesterase Proteins 0.000 description 1
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 101100379081 Emericella variicolor andC gene Proteins 0.000 description 1
- 108010016183 Human immunodeficiency virus 1 p16 protease Proteins 0.000 description 1
- 108090000190 Thrombin Proteins 0.000 description 1
- 102000003990 Urokinase-type plasminogen activator Human genes 0.000 description 1
- 108090000435 Urokinase-type plasminogen activator Proteins 0.000 description 1
- 238000005411 Van der Waals force Methods 0.000 description 1
- 229940022698 acetylcholinesterase Drugs 0.000 description 1
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012362 drug development process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 229910052739 hydrogen Inorganic materials 0.000 description 1
- 239000001257 hydrogen Substances 0.000 description 1
- 125000004435 hydrogen atom Chemical class [H]* 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 239000002547 new drug Substances 0.000 description 1
- 229910052760 oxygen Inorganic materials 0.000 description 1
- 239000001301 oxygen Substances 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 229960004072 thrombin Drugs 0.000 description 1
- 229940043263 traditional drug Drugs 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16C—COMPUTATIONAL CHEMISTRY; CHEMOINFORMATICS; COMPUTATIONAL MATERIALS SCIENCE
- G16C20/00—Chemoinformatics, i.e. ICT specially adapted for the handling of physicochemical or structural data of chemical particles, elements, compounds or mixtures
- G16C20/50—Molecular design, e.g. of drugs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N10/00—Quantum computing, i.e. information processing based on quantum-mechanical phenomena
- G06N10/20—Models of quantum computing, e.g. quantum circuits or universal quantum computers
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Crystallography & Structural Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Pharmacology & Pharmacy (AREA)
- Medicinal Chemistry (AREA)
- General Health & Medical Sciences (AREA)
- Computational Mathematics (AREA)
- Condensed Matter Physics & Semiconductors (AREA)
- Data Mining & Analysis (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Peptides Or Proteins (AREA)
Abstract
The application relates to a molecular docking method and device based on a light quantum computer, belongs to the technical field of drug design, and solves the problems that heuristic algorithms consume long time but cannot obtain standard solutions. The method comprises the following steps: converting the two-dimensional molecular structure diagram of the ligand into a three-dimensional molecular structure diagram; constructing a butt joint box grid point inside a receptor target point based on a three-dimensional molecular structure diagram of the ligand and the receptor so as to optimally match the ligand atom with the receptor grid point; converting the problem of solving the best matching between the ligand atoms and the receptor grid points into searching the maximum weighted group; establishing a secondary unconstrained binary optimization model according to the maximum weighted group, and calculating the strongest binding affinity between the receptor and the ligand under geometric limitation through the secondary unconstrained binary optimization model; and obtaining an optimal solution of the quadratic unconstrained binary optimization model by using the coherent ifer Xin Ji. Solving the molecular modeling problem using the method of I Xin Moxing is more rapid and accurate. The affinity of drugs and proteins can be calculated rapidly using a quantum computer acceleration.
Description
Technical Field
The application relates to the technical field of drug design, in particular to a molecular docking method and device based on a light quantum computer.
Background
Traditional drug screening is a very expensive and resource intensive process, typically costing billions of dollars, with 10% of the power. In recent years, with the development of powerful molecular modeling tools and the increase of the number of protein small molecule complex analytical structures, structure-based drug design is an indispensable tool in new drug development. Molecular docking studies focus on modeling the molecular recognition process by computation. The recognition process aims at simulating the optimal conformation between the protein and the ligand, so that the minimization of the free energy of the whole system is an important link in the early process of drug screening, and the drug development process can be accelerated by means of molecular docking.
Most of screening modes of traditional models in the pharmaceutical field adopt heuristic algorithms and systematic search algorithms, so that the time consumption is long, the optimal solution can not be calculated, and the false positive is high in the drug research and development process. The defects are as follows:
(1) Using a heuristic algorithm: a lot of time is required for iteration, and an optimal solution may not be obtained, so that the calculation amount is large.
(2) The use of systematic search algorithms may face combinatorial explosion problems that are relatively weak with conventional computers.
Conventional models require a large number of samples to achieve a lower energy constellation, and this constellation may not be globally optimal.
Disclosure of Invention
In view of the above analysis, the present application aims to provide a molecular docking method and device based on a light quantum computer, which are used for solving the problems that the heuristic algorithm consumes a long time but cannot obtain a standard solution.
In one aspect, an embodiment of the present application provides a molecular docking method based on a light quantum computer, including: converting the two-dimensional molecular structure diagram of the ligand into a three-dimensional molecular structure diagram; constructing a butt joint box grid point inside a receptor target point based on a three-dimensional molecular structure diagram of the ligand and the receptor so as to optimally match the ligand atom with the receptor grid point; converting solving the best matching problem of the ligand atoms and the receptor grid points into searching a maximum weighted group; establishing a secondary unconstrained binary optimization model according to the maximum weighted group, wherein the strongest binding affinity between the receptor and the ligand under geometric limitation is calculated through the secondary unconstrained binary optimization model; and obtaining an optimal solution of the quadratic unconstrained binary optimization model by using a coherent i Xin Ji.
The beneficial effects of the technical scheme are as follows: the molecular docking method based on the light quantum computer provided by the embodiment of the application solves the molecular simulation problem more quickly and accurately by using the method of I Xin Moxing. The method is accelerated by using a quantum computer, so that the affinity of the medicine and the protein can be rapidly calculated, and researchers are helped to screen out potential lead compounds and assist in the research and development process of the medicine.
Based on a further improvement of the above method, constructing docking box grid points inside a receptor target based on a three-dimensional molecular structure diagram of a ligand and a receptor to optimally match a ligand atom with the receptor grid points comprises: constructing a continuous butt joint box on the receptor structure space according to the size of the ligand; in the grid point matching process, the butt joint boxes of the receptor are scattered into equidistant grid points; placing each atom of a ligand on a grid point to match each atom of the ligand to a receptor grid point; and achieving a receptor-ligand docking pose by minimizing the difference between the matching result and the optimal matching result.
Based on a further improvement of the above method, achieving the receptor-to-ligand docking pose by minimizing the difference between the matching result and the best matching result comprises expressing the ligand-to-receptor docking pose by the following formula:；
wherein,,Rindicating a possible match of the receptor with the ligand,R grid indicating the best match of the receptor to the ligand, | #R-R grid And is the difference distance of the possible match from the best match,meaning that one possible match among all matches is selected such that the difference distance is minimized.
Based on a further improvement of the above method, optimally matching the ligand atoms with the acceptor grid points comprises determining that the matching of two vertices coexist when the following constraint 1 and constraint 2 are satisfied, linking the two vertices: the constraint item 1: the deviation of the distance between the grid points of the receptor docking box and the ligand atoms is smaller than an adjustable threshold parameter; and the constraint term 2: one ligand atom is matched to only one grid point, where each vertex represents one possible match.
Based on a further improvement of the above method, the deviation of the distance between the receptor docking box grid point and the ligand atom being less than the adjustable threshold parameter comprises: for two vertexes in vertex set Va i ,p j ) And%a k ,p l ) The two vertices are edged when the following two constraints are satisfied: i IIr i -r k ‖-‖pr j -pr l ‖|＜c dist Wherein, the method comprises the steps of, wherein,c dist for an adjustable threshold parameter, double vertical lines represent distance, single vertical lines represent absolute values, constraint 1 represents that when the relative distance is coordinated with the ligand shape, both matches coexist; and matching one ligand atom with only one grid point includes:p j ≠p l matching one ligand atom to one grid point; wherein,,A={a i is the set of atoms in the ligand,t i andr i respectively atomsa i Is used for the type and the coordinates of (a),P={p i is a set of butt-box grid points,p i is the coordinates of (a)pr i Vertex set v= {a i ,p j ）|Van（t i ,p j ）＜0}，Van（t i ,p j ) Is of atomic typet i Placed at grid pointsd i Van der Waals energy when above, each vertex represents a possible match.
Based on a further improvement of the above method, a quadratic unconstrained binary optimization model is built from the maximum weighted clique, wherein calculating the strongest binding affinity between the receptor and the ligand under geometrical constraints by means of the quadratic unconstrained binary optimization model comprises calculating energy points by means of the following formula to obtain a scoring function:
；
wherein,,for the decision variable, the value is {0,1}, represent atomsa i Whether or not to be placed at the vertex of the meshd j If yes, 1, otherwise 0;Is the vertex of%a i ,d j ) The second term and the third term in the formula correspond to constraint 1 and constraint 2, respectively, whereinK dist 、K mono Penalty coefficients for two terms, respectively; the first term in the formula means that the more weighted vertices are, the better if #a i ,d j ) And%a k ,d l ) Meeting constraint item 1, then->OtherwiseSimilarly, if%a i ,d j ) And%a k ,d l ) Meeting constraint item 2, then->Otherwise, let(s)>，K dist 、K mono AndC dist three adjustable parameters.
Based on a further improvement of the above method, obtaining the optimal solution of the quadratic unconstrained binary optimization module using coherent i Xin Ji comprises: converting the secondary unconstrained binary optimization module to i Xin Moxing; encoding the i Xin Moxing into a set of amplitude and phase of optical pulses; transmitting the coupling relation between different spins to optical interference by utilizing interference effect; and converting the light pulse into a value of the corresponding spin by using a nonlinear optical effect to obtain an optimal solution of the quadratic unconstrained binary optimization module.
In another aspect, an embodiment of the present application provides a molecular docking device based on a light quantum computer, including: the molecular structure conversion module is used for converting a two-dimensional molecular structure diagram of the ligand into a three-dimensional molecular structure diagram; the matching module is used for constructing a butt joint box grid point in a receptor target point based on a three-dimensional molecular structure diagram of the ligand and the receptor so as to optimally match the ligand atom with the receptor grid point; the problem conversion module is used for converting solving the best matching problem of the ligand atoms and the receptor grid points into searching the maximum weighted group; the model generation module is used for establishing a secondary unconstrained binary optimization model according to the maximum weight group problem, wherein the strongest binding affinity between the receptor and the ligand under the geometric limitation is calculated through the secondary unconstrained binary optimization model; and a model solving module for obtaining an optimal solution of the quadratic unconstrained binary optimization module by using the coherent i Xin Ji.
Based on a further improvement of the above device, the matching module is configured to: the butt joint grid box construction submodule is used for constructing continuous butt joint boxes on the receptor structural space according to the size of the ligand; the discrete submodule is used for dispersing the butt joint boxes of the receptor into equidistant grid points in the grid point matching process; a ligand-matched sub-module for placing each atom of a ligand on a grid point to match each atom of the ligand with a receptor grid point; and a docking sub-module for achieving a docking pose of the receptor and the ligand by minimizing a difference between the matching result and the best matching result.
Based on a further development of the above device, the docking sub-module is configured to express the docking posture of the ligand with the receptor by the following formula:
；
wherein,,Rindicating a possible match of the receptor with the ligand,R grid indicating the best match of the receptor to the ligand, | #R-R grid II is the difference between the possible match and the best matchThe distance between the two adjacent substrates is determined,meaning that one possible match among all matches is selected such that the difference distance is minimized.
Compared with the prior art, the application has at least one of the following beneficial effects:
1. the molecular docking method based on the light quantum computer provided by the embodiment of the application solves the molecular simulation problem more quickly and accurately by using the method of I Xin Moxing.
2. The method is accelerated by using a quantum computer, so that the affinity of the medicine and the protein can be rapidly calculated, and researchers are helped to screen out potential lead compounds and assist in the research and development process of the medicine.
3. The application provides a more excellent model for solving the molecular combination mode by utilizing the characteristics of entangled state, overlapped state and full connection of the quantum computer, and the model is placed at a web end for display and is used for a user.
In the application, the technical schemes can be mutually combined to realize more preferable combination schemes. Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and drawings.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the application, like reference numerals being used to designate like parts throughout the drawings;
FIG. 1 is a flow chart of a molecular docking method based on a light quantum computer according to an embodiment of the present application;
FIG. 2A is a different dimensional structural display of small molecule ligands and receptors according to a first embodiment of the application;
FIG. 2B is a different dimensional structural display of small molecule ligands and receptors according to a second embodiment of the application;
FIG. 2C is a different dimensional structural display of small molecule ligands and receptors according to a third embodiment of the application;
FIG. 2D is a different dimensional structural display of small molecule ligands and receptors according to a fourth embodiment of the application;
FIG. 3 is a schematic diagram of matching acceptor lattice box points to ligand atoms according to an embodiment of the application;
FIG. 4 is a diagram of solving a best matching conversion of a ligand to a maximum clique problem according to an embodiment of the present application;
FIG. 5 is a constraint term schematic of two vertex-added edges in a maximum graph problem diagram in accordance with an embodiment of the present application;
FIG. 6 is an overall flow chart of a molecular docking method based on a light quantum computer according to an embodiment of the present application;
fig. 7 is a block diagram of a molecular docking device based on a light quantum computer according to an embodiment of the present application.
Detailed Description
The following detailed description of preferred embodiments of the application is made in connection with the accompanying drawings, which form a part hereof, and together with the description of the embodiments of the application, are used to explain the principles of the application and are not intended to limit the scope of the application.
Referring to fig. 1, in one embodiment of the present application, a molecular docking method based on a light quantum computer is disclosed, comprising: in step S101, the two-dimensional molecular structure diagram of the ligand is converted into a three-dimensional molecular structure diagram. In a specific embodiment, FIG. 2A shows a ligand (C 13 H 11 N 3 ) A three-dimensional structure of a receptor (thrombin. Alpha.). FIG. 2B shows ligand (C 7 H 9 N 2 ) A two-dimensional structure, a three-dimensional structure of a receptor (urokinase type plasminogen activator). FIG. 2C shows the ligand (C 18 H 19 ClN 2 ) A two-dimensional structure, a three-dimensional structure and a three-dimensional structure of a receptor (acetylcholinesterase). FIG. 2D shows the ligand (C 38 H 40 N 2 O 8 ) Two-dimensional structure, three-dimensional structure and receptor (HIV-1 protease)) Is a three-dimensional structure of (c).
In step S102, docking box grid points inside the receptor target are constructed based on the three-dimensional molecular structure diagram of the ligand and the receptor to optimally match the ligand atoms with the receptor grid points. Referring to fig. 3, constructing docking box grid points inside a receptor target based on a three-dimensional molecular structure diagram of a ligand and a receptor to optimally match a ligand atom to the receptor grid points comprises: constructing continuous butt-joint boxes on the receptor structure space according to the ligand size and dispersing the butt-joint boxes into equidistant grid points; in the grid point matching process, the grid points of the butt joint box of the receptor are scattered into equidistant grid points; placing each atom of the ligand on a grid point to match each atom of the ligand with a receptor grid point; and achieving a receptor-ligand docking pose by minimizing the difference between the matching result and the optimal matching result.
Achieving a receptor-to-ligand docking pose by minimizing the difference between the matching result and the best matching result includes formulating the ligand-to-receptor docking pose by the following equation:
；
wherein,,Rindicating a possible match of the receptor with the ligand,R grid indicating the best match of the receptor to the ligand, | #R-R grid And is the difference distance between the possible match and the best match,meaning that one possible match among all matches is selected such that the difference distance is minimized.
Referring to fig. 4, optimally matching a ligand atom with a receptor grid point includes determining that matching of two vertices coexist, linking the two vertices, and matching when the following constraint 1 and constraint 2 are satisfied: constraint item 1: the deviation of the distance between the grid points of the receptor docking box and the ligand atoms is smaller than an adjustable threshold parameter; constraint term 2: one ligand atom is matched to only one grid point, where each vertex represents one possible match.
Constraint item 1: the deviation of the distance between the grid points of the receptor docking box and the ligand atoms is smaller than an adjustable threshold value parameter, which comprises the following steps ofa i ,p j ) And%a k ,p l ) The two vertices are bordered when the following two constraints are satisfied:
|‖r i -r k ‖-‖pr j -pr l ‖|＜c dist ,
wherein,,c dist for an adjustable threshold parameter, double vertical lines represent distance, single vertical lines represent absolute values, constraint 1 represents that both matches coexist when the relative distance is coordinated with the ligand shape.A={a i Is the set of atoms in the ligand,t i andr i respectively atomsa i Is used for the type and the coordinates of (a),P={p i is the butt-joint box grid point,p i is the coordinates of (a)pr i Vertex set v= {a i ,p j ）|Van（t i ,p j ）＜0}，Van（t i ,p j ) Is of atomic typet i Placed at grid pointsd i Van der Waals energy when above, each vertex represents a possible match. Constraint term 2: matching one ligand atom to only one grid point includes:p j ≠p l so that one ligand atom is matched to one grid point.
In step S103, the problem of solving the best match of ligand atoms to receptor grid points is converted into a problem of finding the largest weighted group. Solving for the ligand grid point best matching transformation is done to find the largest weight group of G (V, E) in order to find more powerful protein-ligand interactions under geometric constraints.
In step S104, a quadratic unconstrained binary optimization model is built from the maximum weighted group, wherein the strongest binding affinity between the receptor and the ligand under geometrical constraints is calculated by the quadratic unconstrained binary optimization model. Establishing a quadratic unconstrained binary optimization model from the maximum weighted mass, wherein calculating the strongest binding affinity between the receptor and the ligand under geometric constraints by the quadratic unconstrained binary optimization model comprises calculating energy points to obtain scoring functions by the following formula:
；
wherein,,for decision variables, the value is {0,1}, representing an atoma i Whether or not to be placed at the vertex of the meshd j If yes, 1, otherwise 0;Is the vertex of%a i ,d j ) The second term and the third term in the formula correspond to constraint 1 and constraint 2, respectively, whereinK dist 、K mono Penalty coefficients for two terms, respectively; the first term in the formula means that the more weighted vertices are, the better if #a i ,d j ) And%a k ,d l ) Meeting constraint item 1, then->OtherwiseSimilarly, if%a i ,d j ) And%a k ,d l ) Meeting constraint item 2, then->Otherwise, let(s)>，K dist 、K mono AndC dist three adjustable parameters.
In step S105, an optimal solution of the quadratic unconstrained binary optimization model is obtained using the coherent i Xin Ji. The obtaining of the optimal solution of the quadratic unconstrained binary optimization module by using the coherent i Xin Ji comprises: converting the secondary unconstrained binary optimization module into an I Xin Moxing; encoding i Xin Moxing into the amplitude and phase of a set of optical pulses; transmitting the coupling relation between different spins to optical interference by utilizing interference effect; and converting the light pulse into a value corresponding to the spin by using a nonlinear optical effect to obtain an optimal solution of the quadratic unconstrained binary optimization module.
Compared with the prior art, the molecular docking method based on the light quantum computer provided by the embodiment is faster and more accurate in solving the molecular simulation problem by using the method of the Xin Moxing. The method is accelerated by using a quantum computer, so that the affinity of the medicine and the protein can be rapidly calculated, and researchers are helped to screen out potential lead compounds and assist in the research and development process of the medicine.
Hereinafter, a molecular docking method based on a light quantum computer according to an embodiment of the present application will be described in detail by way of specific examples with reference to fig. 2A to 6.
Molecular docking: the complex mode is found by calculating the space complementation and energy matching between ligand receptors based on the lock key model of ligand receptor identification, and the complex mode is widely applied in the field of drug design and can be divided into rigid butt joint, semi-flexible butt joint and flexible butt joint.
The molecular docking method based on the light quantum computer can be used for drug development. The method can screen a huge amount of medicine libraries with high flux, a user inputs a molecular structure and a protein structure, and the server can calculate the affinity of the medicine libraries and return the affinity to the user, so that the medicine research and development process is accelerated. The application uses a photon computer to convert the important molecular docking problem in the drug screening process into a QUBO model, converts a molecular three-dimensional model into a mathematical graph, uses the graph form to represent the ligand and the receptor, and realizes the prediction of the ligand receptor binding mode and the evaluation of the affinity thereof through the QUBO model by mathematical modeling. The user can upload different small molecules and protein structures, the subsequent server can convert the small molecules and the protein structures, an optimal calculation result is given out through the light quantum computer, and a three-dimensional model is displayed for the user. For the drug screening problem, the light quantum computer can give the result faster and the result is more accurate than the traditional computer.
1. Molecular display
First, the three-dimensional structures of the molecules obtained by the application are all from Protein Data Bank and the PDB database (adopting crystal structures or structures after energy minimization), and fig. 2A, 2B, 2C and 2D show two-dimensional and three-dimensional structures of four ligands, and then the 3D structures of the molecules are simplified into the form of mathematical medium graphs: the simplified structure is shown as a graph weighted by distance matrix/edge.
2. Construction of Butt-joint grid Box (GridBox) Point inside receptor target
Referring to fig. 3, a continuous docking box is first constructed on the receptor structure space according to the size of the ligand, and in Grid point matching, the receptor docking box is first discretized into Grid points with a spacing of 2a, and then each atom of the ligand is placed on one Grid point, i.e., each atom of the ligand is matched with the receptor Grid point. The true docking posture is:
；
wherein the method comprises the steps ofRIs an arbitrary solution in all of the solution spaces,R grid corresponding to the best match IIR-R grid II is solutionRThe difference distance from the best match is calculated by minimizing the solutionRThe difference from the best match achieves a true docking pose. Such minimization may be achieved by a Kabsch RMSD rotation matrix (Prody). The docking box grid points inside the receptor targets are generated by autoprid in the ADFR package.
3. Converting solving the best match of the ligands to solve the maximum-clique problem (Maximum Clique Problem, MCP)
In the grid point matching solution problem, the set of all atomic compositions in the ligand isA={a i }，t i Andr i is thata i Type (e.g., hydrogen, carbon, oxygen) and coordinates. The combination of the grid point components in all the butt joint boxes isP={p i Apexes in squarep i The corresponding coordinates arepr i . Here we convert solving the receptor and ligand best matching problem to one solving the biggest cluster problem (FIG. 4). The specific transformation process is as follows: given a pictureG(V,E) In, vertex setV={（a i ,p j ）|Van（t i ,p j ) < 0}, whereina i Represents a ligand atom which is a group of atoms,p j representing the vertex of the docking box after discretization of the receptor,Van（t i ,p j ) Is an atoma i The corresponding atom typet i Placed at receptor grid pointsd j Van der Waals force at the top and vertex [ ]a i ,p j ) Is given by the weight of. Each vertex here represents a possible match, the weight of which shows its contribution to the overall binding affinity. For two vertexes in Va i ,p j ) And%a k ,p l ) The two vertices are bordered when the following two constraints are satisfied: constraint 1 is ||r i -r k ‖-‖pr j -pr l ‖|＜c dist Here, wherec dist Is a threshold value, is a singleAdjusting parameters, wherein double vertical lines represent distances, single vertical lines represent absolute values, constraint 1 represents that when the relative distances are coordinated with the ligand shape, two kinds of matching can coexist; constraint 2 isp j ≠p l When it is indicated that an atom can only be matched with a grid point. Based on the above description, solving for ligand-grid point best match conversion to findG(V,E) To find more and stronger protein-ligand interactions under geometric constraints.
4. Mathematical model
After the above process is completed, we finally construct a Graph (Graph) with weights on the nodesG(V,E). The 3d gestures and scoring function information are stored in the nodes and edges of the graph. The maximum weight cluster problem is the NP-Hard problem, which is difficult for conventional computers to give solutions. Here we convert it to the QUBO formula (formula 1). In the drawingsG(V,E)Each vertex represents a possible match. The weight of a vertex is defined as the contribution of the ligand atom and receptor to the box grid point that matches the binding affinity. Only when a match of two vertices can coexist will the two vertices be bordered, i.e. connected, as shown in FIG. 5. On the one hand, they should adhere to geometric constraints to maintain the shape of the ligand. The deviation of the distance between the spatial point and the ligand atom should be below a critical valuec dist (constraint 1). On the other hand, a ligand atom can only be matched to one spatial point (constraint 2). Following the above definition, the graphThe largest weighted group of (2) represents the strongest binding affinity between protein and ligand under geometric constraints and is also a good docking pose.
FIG. 5 is a schematic illustration of constraint terms intended to determine a constructed graphG(V,E)Two points of the Chinese medicinea i ,p j ) And%a k ,p l ) Whether or not there is an edge between. Edges are added when the following two constraints are satisfied: 1. if it is%a i ,p j ) And any other pointa k ,p l ) Distance from the original liganda i Atoms anda k the distance between atoms is less than a thresholdC dist 。2.i≠jI.e. one atom does not appear at multiple lattice pointsd i ，If the constraint condition is violated, the edges of the two points do not exist. The best match to solve the ligand is converted to a maximum cluster problem result and the point-to-point pairing is obtained after the maximum weighted cluster is found. We use the Kabsch RMSD rotation matrix to superimpose atomic coordinates on the grid points of the receptor docking box after discretization, and then generate the final docking pose. Finally, we can convert the atom-atom pairs into 3D coordinates. And the energy points can be calculated to obtain the final scoring function.
；
Wherein,,for decision variables, the value is {0,1}, representing an atoma i Whether or not to be placed at the vertex of the meshd j If yes, 1, otherwise 0;Is the vertex of%a i ,d j ) Is a weight of (2). The latter two terms in the formula correspond to two constraint terms 1 and 2 in 3, respectively, whereK dist 、K mono Penalty coefficients for the two terms, respectively. The first term means that the more weighted vertices the better. If it is%a i ,d j ) And%a k ,d l ) Meeting constraint 1, then->Otherwise->. Similarly, if%a i ,d j ) And%a k ,d l ) Meeting constraint 2, then->. Otherwise the first set of parameters is selected,。K dist 、K mono andC dist is three adjustable parameters.
The model is a quadratic unconstrained binary optimization model (Quadratic Unconstrained Binary Optimization, QUBO) which can be solved on the model i Xin Xianggan i Xin Ji (coherent Ising machine, CIM) after being converted into an Ising model. Each binary variable is represented by the state of one qubit in the CIM, while hamilton corresponds to the energy of the quantum system. CIM uses nonlinear optical effects based on optical interference to achieve solution of the QUBO problem. Specifically, QUBO is problem-translated into a special Ising model, which is then encoded by the CIM into the amplitude and phase of a set of optical pulses, which imparts the coupling relationship between the different spins to the optical interference using interference effects. And then, converting the light pulse into a value corresponding to spin by using a nonlinear optical effect, and finally obtaining the ground state of the Ising model, namely the optimal solution of the QUBO problem. Its advantages are high effect on handling the large-scale QUBO problem, high parallelism and high response speed. Thus, the best solution can be read from the ground state of the qubit.
5. Test performance of the model on the CASF-2016 core set data set
Based on the mathematical model, a ligand matching algorithm is constructed and solved by using a CIM quantum computer, and three optimal parameters of the model are obtained through training on a training data set:Cdist = 2.565 ,K dist = 2.337,K mono = 39.530, the model was then tested on the CASF-2016 core set dataset under the parameters described aboveThe CASF-2016 core set is a relatively small high quality protein-ligand complex dataset that is commonly used to verify docking methods. Root mean square deviation (Root Mean Square Deviation, RMSD) is often used to measure the degree of atomic deviation from aligned positions, and by calculating this value, the quality of matching of the atomic positions of the ligands can be determined. Through testing, 257 examples of the method can achieve matching, and 75% of ligand complexes RMSD in a successfully paired ligand set are smaller than 2a (table 1), which indicates that the model can achieve matching of ligands better.
6. Flow chart
Referring to fig. 6, when a user uses the method, only the screened drug library file and the target protein structure file need to be input, the initial ligand structure and the receptor structure of the drug library file are displayed at the user side, the user can conveniently perform visual three-dimensional operation, then after the target range is determined, data are transmitted back to the server, the server builds a QUBO model according to the parameters and transmits the QUBO model to the light quantum computer, an optimal solution is calculated, time consumption is greatly reduced under the condition that the quantum computer is used in the whole process, and the user can see the screened structure model and score at the use interface. Subsequent experimental verification was performed using these molecules to obtain the lead compound.
Referring to fig. 7, another embodiment of the present application discloses a molecular docking device based on an optical quantum computer, which includes a molecular structure conversion module 701, a matching module 702, a problem conversion module 703, a model generation module 704, and a model solving module 705.
The molecular structure conversion module 701 is configured to convert a two-dimensional molecular structure diagram of a ligand into a three-dimensional molecular structure diagram.
The matching module 702 is configured to construct docking box grid points within a receptor target based on a three-dimensional molecular structure diagram of the ligand and the receptor to best match the ligand atoms to the receptor grid points. The matching module 702 includes: the butt joint grid box construction submodule is used for constructing continuous butt joint boxes on the receptor structural space according to the size of the ligand; the discrete submodule is used for dispersing the butt joint boxes of the receptor into equidistant grid points in the grid point matching process; a ligand-matched sub-module for placing each atom of the ligand on a grid point to match each atom of the ligand with a receptor grid point; and a docking sub-module for achieving a docking pose of the receptor and the ligand by minimizing a difference between the matching result and the best matching result.
The docking sub-module is used to express the docking pose of the ligand to the receptor by the following formula:
the method comprises the steps of carrying out a first treatment on the surface of the Wherein,,Rrepresents one possible match of receptor to ligand,R grid indicating the best match of the receptor to the ligand, | #R-R grid II is the difference distance between the possible match and the best match, +.>Meaning that one possible match among all matches is selected such that the difference distance is minimized. />
The problem transformation module 703 is used to transform the best matching problem solving the ligand atoms and acceptor grid points into finding the largest weighted group.
The model generation module 704 is configured to build a quadratic unconstrained binary optimization model according to the maximum weight group problem, wherein the strongest binding affinity between the receptor and the ligand under geometric constraints is calculated by the quadratic unconstrained binary optimization model.
The model solving module 705 is configured to obtain an optimal solution of the quadratic unconstrained binary optimization module by using the coherent i Xin Ji.
Those skilled in the art will appreciate that all or part of the flow of the methods of the embodiments described above may be accomplished by way of a computer program to instruct associated hardware, where the program may be stored on a computer readable storage medium. Wherein the computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, etc.
The present application is not limited to the above-mentioned embodiments, and any changes or substitutions that can be easily understood by those skilled in the art within the technical scope of the present application are intended to be included in the scope of the present application.
Claims (7)
1. The molecular docking method based on the light quantum computer is characterized by comprising the following steps of:
converting the two-dimensional molecular structure diagram of the ligand into a three-dimensional molecular structure diagram;
constructing receptor grid points inside a receptor target point based on a three-dimensional molecular structure diagram of the ligand and the receptor so as to optimally match the ligand atoms with the receptor grid points;
converting solving the best matching problem of the ligand atoms and the receptor grid points into searching a maximum weighted group;
establishing a secondary unconstrained binary optimization model according to the maximum weighted group, wherein the strongest binding affinity between the receptor and the ligand under geometric limitation is calculated through the secondary unconstrained binary optimization model; and
obtaining an optimal solution of the secondary unconstrained binary optimization model by using a coherent i Xin Ji;
the method further comprises the step of establishing a quadratic unconstrained binary optimization model according to the maximum weighted group, wherein calculating the strongest binding affinity between the receptor and the ligand under geometrical constraints through the quadratic unconstrained binary optimization model comprises calculating energy points through the following formula to obtain a scoring function:
；
wherein, H is a scoring function,for decision variables, the value {0,1} represents the ligand atom a i Whether or not to be placed at receptor grid point d j If yes, 1, otherwise 0;Is the vertex (a) i ,d j ) The second and third terms in the formula correspond to constraint 1 and constraint 2, respectively, where K dist 、K mono Penalty coefficients for two terms, respectively; if (a) i ,d j ) And (a) k ,d l ) Conforming to constraint 1, then
Otherwise->Similarly, if (a) i ,d j ) And (a) k ,d l ) Meeting constraint item 2, then->Otherwise, let(s)>The method comprises the steps of carrying out a first treatment on the surface of the The best matching of ligand atoms to receptor grid points includes: for two vertices (a i ,p j ) And (a) k ,p l ) When the following constraint item 1 and constraint item 2 are satisfied, it is determined that matching of two vertices coexist, and the two vertices are connected: the constraint item 1: the deviation in distance between the receptor grid points and the ligand atoms is less than an adjustable threshold parameter, comprising: i II r i -r k ‖-‖pr j -pr l ‖|＜c dist Wherein c dist For the adjustable threshold parameter, the double vertical lines represent distances, the single vertical line represents absolute values, and the constraint term 1 represents that when the relative distance is coordinated with the ligand shape, two matches coexist;
the constraint term 2: one ligand atom matches only one acceptor lattice point, comprising: p is p j ≠p l Such that one ligand atom matches one receptor lattice point;
wherein a= { a i And } is the set of atoms in the ligand, t i And r i Respectively ligand atom a i P= { P i Is the receptor grid point set, p i Is pr as the coordinate i Vertex set v= { (a) i ,p j ）|Van（t i ,p j ）＜0}，Van（t i ,p j ) Is of type t i Is placed at receptor lattice point p j Van der Waals energy at the time of the upper.
2. The quantum computer-based molecular docking method of claim 1, wherein constructing receptor lattice points within a receptor target based on a three-dimensional molecular structure diagram of a ligand and a receptor to optimally match a ligand atom to the receptor lattice points comprises:
constructing a continuous butt joint grid box on the receptor structure space according to the size of the ligand;
in the grid point matching process, the butt joint grid boxes of the receptor are discretized into equidistant grid points;
placing each atom of a ligand on one receptor grid point to match each atom of the ligand to the receptor grid point; and
docking of the receptor with the ligand is achieved by minimizing the difference between the matching result and the best matching result.
3. The molecular docking method based on a light quantum computer of claim 2, wherein the achieving a docking gesture of the receptor with the ligand by minimizing a difference between the matching result and the optimal matching result comprises:
the docking position of the ligand to the receptor is expressed by the following formula:
；
wherein R represents a possible match of the receptor with the ligand, R grid Represents the best match of the receptor to the ligand, |R-R grid And is the difference distance of the possible match from the best match,meaning that one possible match among all matches is selected such that the difference distance is minimized.
4. The method of molecular docking based on an optical quantum computer of claim 1, wherein obtaining an optimal solution of the quadratic unconstrained binary optimization model using coherent i Xin Ji comprises:
converting the quadratic unconstrained binary optimization model to i Xin Moxing;
encoding the i Xin Moxing into a set of amplitude and phase of optical pulses;
transmitting the coupling relation between different spins to optical interference by utilizing interference effect; and
and converting the optical pulse into a value of the corresponding spin by using a nonlinear optical effect to obtain an optimal solution of the quadratic unconstrained binary optimization model.
5. A molecular docking device based on a light quantum computer for implementing the method of any one of claims 1-4, comprising:
the molecular structure conversion module is used for converting a two-dimensional molecular structure diagram of the ligand into a three-dimensional molecular structure diagram;
the matching module is used for constructing receptor grid points in the receptor target based on a three-dimensional molecular structure diagram of the ligand and the receptor so as to optimally match the ligand atoms with the receptor grid points;
the problem conversion module is used for converting solving the best matching problem of the ligand atoms and the receptor grid points into searching the maximum weighted group;
the model generation module is used for establishing a secondary unconstrained binary optimization model according to the maximum weighted group, wherein the strongest binding affinity between the receptor and the ligand under the geometric limitation is calculated through the secondary unconstrained binary optimization model; and
and the model solving module is used for acquiring an optimal solution of the quadratic unconstrained binary optimization model by using the coherent I Xin Ji.
6. The molecular docking device based on an optical quantum computer of claim 5, wherein the matching module is configured to:
the butt joint grid box construction submodule is used for constructing continuous butt joint grid boxes on the receptor structure space according to the size of the ligand;
the discrete submodule is used for dispersing the butt joint grid boxes of the receptor into equidistant grid points in the grid point matching process;
a ligand-matched sub-module for placing each atom of a ligand on one receptor grid point to match each atom of the ligand with the receptor grid point; and
and a docking sub-module for achieving a docking posture of the receptor and the ligand by minimizing a difference between the matching result and the optimal matching result.
7. The molecular docking device of claim 6, wherein the docking sub-module is configured to express the docking pose of the ligand and the receptor by the following formula:
；
wherein R represents a possible match of the receptor with the ligand, R grid Represents the best match of the receptor to the ligand, |R-R grid And is the difference distance of the possible match from the best match,meaning that one possible match among all matches is selected such that the difference distance is minimized.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310882617.XA CN116597917B (en) | 2023-07-19 | 2023-07-19 | Molecular docking method and device based on light quantum computer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310882617.XA CN116597917B (en) | 2023-07-19 | 2023-07-19 | Molecular docking method and device based on light quantum computer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116597917A CN116597917A (en) | 2023-08-15 |
| CN116597917B true CN116597917B (en) | 2023-10-20 |
Family
ID=87606658
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310882617.XA Active CN116597917B (en) | 2023-07-19 | 2023-07-19 | Molecular docking method and device based on light quantum computer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116597917B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117423379B (en) * | 2023-12-19 | 2024-03-15 | 合肥微观纪元数字科技有限公司 | Molecular screening method and related device adopting quantum computation |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114678082A (en) * | 2022-03-08 | 2022-06-28 | 南昌立德生物技术有限公司 | Computer-aided virtual high-throughput screening algorithm |
| CN114882940A (en) * | 2022-03-28 | 2022-08-09 | 北京玻色量子科技有限公司 | Molecular docking method and device based on coherent Icin machine |
| CN116052802A (en) * | 2023-03-31 | 2023-05-02 | 北京玻色量子科技有限公司 | Coherent Yi Xin Ji, polypeptide design method and device based on coherent Yi Xin Ji |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010036927A2 (en) * | 2008-09-26 | 2010-04-01 | The Regents Of The University Of Michigan | High-throughput ensemble-based docking and elucidation of 3-dimensional structural conformations of flexible biomolecular targets |
| CA2769045A1 (en) * | 2009-05-04 | 2010-11-11 | University Of Maryland, Baltimore | Method for binding site identification by molecular dynamics simulation (silcs:site identification by ligand competitive saturation) |
-
2023
- 2023-07-19 CN CN202310882617.XA patent/CN116597917B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114678082A (en) * | 2022-03-08 | 2022-06-28 | 南昌立德生物技术有限公司 | Computer-aided virtual high-throughput screening algorithm |
| CN114882940A (en) * | 2022-03-28 | 2022-08-09 | 北京玻色量子科技有限公司 | Molecular docking method and device based on coherent Icin machine |
| CN116052802A (en) * | 2023-03-31 | 2023-05-02 | 北京玻色量子科技有限公司 | Coherent Yi Xin Ji, polypeptide design method and device based on coherent Yi Xin Ji |
Non-Patent Citations (1)
| Title |
|---|
| 蛋白质分子柔性对接方法研究进展;王乐希;郑珩;劳兴珍;;药物生物技术(第04期);72-76 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116597917A (en) | 2023-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Krone et al. | Visual analysis of biomolecular cavities: State of the art | |
| Anwer et al. | From solid modelling to skin model shapes: Shifting paradigms in computer-aided tolerancing | |
| Khan et al. | Surface remeshing: A systematic literature review of methods and research directions | |
| Freytag et al. | Finite element analysis in situ | |
| CN116597917B (en) | Molecular docking method and device based on light quantum computer | |
| Kozlikova et al. | Visualization of biomolecular structures: State of the art | |
| CN114882940B (en) | Molecular docking method and device based on coherent Icin machine | |
| KR20070026497A (en) | Calculating three-dimensional (3d) voronoi diagrams | |
| Anderson et al. | Smooth, volume-accurate material interface reconstruction | |
| Krone et al. | Interactive extraction and tracking of biomolecular surface features | |
| Maquart et al. | 3D volumetric isotopological meshing for finite element and isogeometric based reduced order modeling | |
| Simões et al. | CavVis—a field-of-view geometric algorithm for protein cavity detection | |
| Hermosilla et al. | Interactive gpu-based generation of solvent-excluded surfaces | |
| Liu et al. | Parameterization for molecular Gaussian surface and a comparison study of surface mesh generation | |
| Li et al. | Meshless harmonic volumetric mapping using fundamental solution methods | |
| Xu et al. | Developing an extended IFC data schema and mesh generation framework for finite element modeling | |
| Ataei et al. | XLB: A differentiable massively parallel lattice Boltzmann library in Python | |
| Cerutti et al. | DRACO-STEM: an automatic tool to generate high-quality 3d meshes of shoot apical meristem tissue at cell resolution | |
| Kim et al. | Interaction interfaces in proteins via the Voronoi diagram of atoms | |
| Li et al. | Interactive deformation and cutting simulation directly using patient‐specific volumetric images | |
| CN116994660A (en) | Method, device, equipment and storage medium for generating composite structure | |
| Chang et al. | Review of Three-Dimensional Model Simplification Algorithms Based on Quadric Error Metrics and Bibliometric Analysis by Knowledge Map | |
| Qin et al. | Integrated modeling, simulation, and visualization for nanomaterials | |
| Xu et al. | Surface-based Multimodal Protein-Ligand Binding Affinity Prediction | |
| Behar et al. | Dynamic Minkowski sums under scaling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |